Closing Case IBM's Watson

The Problem

Computer scientists have long sought to design computer-based information systems (CBISs) that possess general problem-solving skills. That is, researchers want CBIS to interact in natural human terms across a range of applications and processes, comprehending the questions that humans ask and providing answers that humans can understand. An important step toward this goal is the development of the IBM Watson system.

An Interesting IT Solution

IBM (www.ibm.com) has developed an artificial intelligence (we discuss artificial intelligence in Technology Guide 4) CBIS capable of answering questions posed in natural language. IBM named the system "Watson," after the company's founder, Thomas J. Watson, Sr. IBM describes Watson as an application of advanced natural language processing, information retrieval, knowledge representation and reasoning, and machine learning technologies to the field of open domain (general) question answering.
IBM developed Watson specifically to answer questions on the quiz show Jeopardy!. In 2011, Watson competed on Jeopardy against former winners Brad Rutter and Ken Jennings. Watson won the game series and received the first prize of $1 million. (In Jeopardy! the host reads the answer, and the contestants must then provide the correct question.)

For Jeopardy!, Watson had access to 200 million pages of structured and unstructured content consuming 4 terabytes of disk storage, including the full text of Wikipedia (see www .wikipedia.org). Signify cantle, however, Watson was not connected to the Internet during the game. For each clue, Watson's three most probable responses were displayed on the television screen. The probabilities represented Watson's level of confidence in any given answer. Watson consistently outperformed its human opponents, but it did experience problems responding to a few categories, notably those having short clues containing only a few words.
Following the television performance, IBM executives immediately turned their attention to commercializing Watson.

They decided to focus on a fi eld in which Watson could have a distinctive social impact while also proving its ability to master a complex body of knowledge. The team chose medicine.

Medicine was a logical choice. Although some health information is structured-for example, blood pressure readings and cholesterol counts-the vast majority is unstructured. This information includes textbooks, medical journals, patient records, and nurse and doctor notes. In fact, modern medicine entails so much unstructured information that its rapid growth has surpassed the ability of practitioners to keep up. Keep in mind that IBM has made it clear that Watson is not intended to replace doctors. Rather, its purpose is to assist them in avoiding medical errors and sharpening their medical diagnoses. Enter Watson. Watson can ingest more data in one day that any human could in a lifetime. It can read all of the world's medical journals in seconds. Simultaneously, it can read patient histories, monitor the latest drug trials, examine the potency of new therapies, and closely follow state-of-the-art guidelines that help doctors choose the best treatments. Watson can also analyze images such as MRIs and EKGs. Watson never goes on vacation; never forgets a fact (in fact, Watson keeps on learning); never gets tired; and works 24/7/365.
The Results: Initial Uses of Watson

Watson at WellPoint. IBM partnered with WellPoint (www .wellpoint.com), the largest managed for-profi t healthcare company in the Blue Cross and Blue Shield association, to increase the effi ciency of health insurance decisions and claims by utilizing Watson's analytic capabilities. At Well- Point, Watson was trained with more than 25,000 test case scenarios and 1,500 real-life cases. It acquired the ability to analyze queries in the context of complex medical data and human natural language, including doctor's notes, patient records, and medical annotations and clinical feedback. In addition, nurses spent more than 14,700 hours of hands-on training with Watson. Watson continues to learn on the job, while working with the WellPoint nurses who originally conducted its training.
In December 2012, Watson started processing common medical procedure requests by physicians for members in WellPoint-affiliated health plans in five physician offices in the Midwest. The overall goal is for Watson to streamline the review processes between a patient's physician and his or her health plan.

Watson at Memorial Sloan-Kettering. IBM also sent Watson to the Memorial Sloan-Kettering Cancer Center (MSK; www.mskcc.org) in Manhattan to help physicians diagnose and treat lung cancer. Watson's role is to analyze a patient's specifi c symptoms, medical history, and hereditary history, and then synthesize those data with available unstructured and structured medical information, including published medical books and articles. By February 2013, Watson had ingested more than 600,000 pieces of medical information and 2 million pages of text from 42 medical journals and clinical trials in the area of oncology research. (Oncology is the study and treatment of cancer.) MSK has immersed Watson in the complexities of cancer as well as the explosion of genetic research that has set the stage for changing treatment practices for many cancer patients by basing their treatments on their personal genetic profi les. Starting with 1,500 actual lung cancer cases, MSK clinicians are training Watson to extract and interpret physician notes, lab results, and clinical research, while sharing its expertise with other physicians.

Watson at the Cleveland Clinic. The Cleveland Clinic Lerner College of Medicine of Case Western Reserve University is continuing Watson's medical training by increasing its health expertise. As with other installations of Watson, the goal is for Watson to assist medical professionals in diagnosing and treating patients.
Watson at MCCM and Westmed. In February 2013, IBM announced that oncologists at the Maine Center for Cancer Medicine (www.mccm.org) and Westmed Medical Group (www.westmedgroup.com) in New York had started to use Watson to help diagnose lung cancer and recommend treatment.

Watson at Rensselaer. On January 30, 2013, the Rensselaer Polytechnic Institute (RPI) (www.rpi.edu) received Watson. Watson will be housed at the Institute's technology park and will be available to researchers and students.
Watson at Citigroup. IBM intends to use Watson in other information-intensive fi elds, such as telecommunications, financial services, and government. At Citigroup (www.citigroup.com), for example, Watson analyzes financial, regulatory, economic, and social data across fi nancial exchanges, currencies, and funds to help simplify and improve the bank's digital interactions with its customers. One last interesting point: In March 2013, IBM executives noted that Watson's performance had improved by 240 percent since its Jeopardy! appearance. Buoyed by this success, IBM is now planning for "Watson 2.0," which will be suffi iently energy efficient to function on smartphones and tablet computers.

Questions
1. What applications can you think of for Watson in a university setting?
2. What are potential disadvantages of using Watson in healthcare settings?
3. Would you consider being diagnosed only by Watson? Why or why not?
4. Would you consider being diagnosed by your personal physician, if he or she consulted Watson? Why or why not?